For the time series of development, I use the tracker from [Deutsche Welle](https://p.dw.com/p/3lUln), with  of vaccines trials and their trial start and (tentative) completion dates.
Records include three phases: I, II, and III. No pre-trials are recorded. Some trials are recorded as "dual phases" (e.g. Phase I/II), and I reassign them to the higest
(e.g. Phase I/II to Phase II). Vaccines in multiple phases in parallel are counted separately.

For the trial *development* time series, I set a dummy to be 1 for dates between the start and completion date for each , 0 otherwise, and then sum over all vaccines. For the trial *completed* time series, I set a dummy to be 1 for dates after the completion date for each , 0 otherwise, and then sum over all vaccines. 
